Wheel rail damage detection method and device, electronic equipment and storage medium

Abstract

The invention discloses a wheel rail damage detection method and device, electronic equipment and a storage medium. The method comprises the following steps: acquiring target signal data when a vehicle passes through a target wheel track; performing matrix decomposition processing of a preset mode on the target signal data to obtain at least one decomposition vector; according to the gravity center frequency of each decomposition vector, reconstructing each decomposition vector to obtain reconstructed signal data; performing preset modal decomposition processing on the reconstructed signal data to obtain at least one modal component; dividing each modal component into two groups according to a preset frequency threshold, and calculating an energy ratio of the two groups of modal components; and determining that a flaw signal exists in the target signal data according to the energy ratio. According to the technical scheme provided by the embodiment of the invention, the identification efficiency and the identification accuracy of the damage signal of the wheel track are improved.

Description

Technical Field

[0001] The present application relates to the field of track detection technology, and in particular to a wheel-rail damage detection method, device, electronic equipment and storage medium. Background Art

[0002] With the continuous development of transportation technology, railways have become an indispensable form of transportation in production and life. In order to ensure the safety of railway transportation, wheel and rail damage detection has become a top priority in railway technology.

[0003] Currently, there are two main methods for detecting damage or defects in wheels and rails. One is manual inspection, where maintenance personnel use flaw detection equipment to conduct inspections during low-speed periods. However, due to the large area of ​​railway paving, the detection efficiency is low. The other is an acoustic detection method. Since guided waves will reflect when encountering defects during their propagation on the wheels and rails, damage detection is performed on the rails by extracting the reflected echoes. However, due to the complex signal noise propagating on the wheels and rails, the efficiency and accuracy of identifying damage need to be improved. Summary of the Invention

[0004] The present application provides a wheel-rail damage detection method, device, electronic device and storage medium to improve the recognition efficiency and accuracy of wheel-rail damage signals.

[0005] According to one aspect of the present application, a wheel-rail damage detection method is provided, comprising:

[0006] Obtain target signal data when a vehicle passes on the target wheel track;

[0007] Performing a matrix decomposition process on the target signal data in a preset manner to obtain at least one decomposition vector;

[0008] Reconstruct each decomposition vector according to the centroid frequency of each decomposition vector to obtain reconstructed signal data;

[0009] Performing a preset modal decomposition process on the reconstructed signal data to obtain at least one modal component;

[0010] Each modal component is divided into two groups according to a preset frequency threshold, and the energy ratio of the two groups of modal components is calculated;

[0011] According to the energy ratio, it is determined that there is a damaged signal in the target signal data.

[0012] According to another aspect of the present application, a wheel-rail damage detection device is provided, comprising:

[0013] A signal acquisition module is used to obtain target signal data when a vehicle passes on the target wheel track;

[0014] A matrix decomposition module, configured to perform a matrix decomposition process on the target signal data in a preset manner to obtain at least one decomposition vector;

[0015] A signal reconstruction module is used to reconstruct each decomposition vector according to the centroid frequency of each decomposition vector to obtain reconstructed signal data;

[0016] A modal decomposition module, configured to perform a preset modal decomposition process on the reconstructed signal data to obtain at least one modal component;

[0017] An energy comparison module is used to divide each modal component into two groups according to a preset frequency threshold and calculate the energy ratio of the two groups of modal components;

[0018] The damage determination module is used to determine whether a damage signal exists in the target signal data based on the energy ratio.

[0019] According to another aspect of the present application, an electronic device is provided, comprising:

[0020] at least one processor; and

[0021] a memory communicatively connected to the at least one processor; wherein,

[0022] The memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the wheel-rail damage detection method described in any embodiment of the present application.

[0023] According to another aspect of the present application, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the wheel-rail damage detection method described in any embodiment of the present application when executed.

[0024] According to another aspect of the present application, a computer program product is provided, which includes a computer program, and when the computer program is executed by a processor, it implements the wheel-rail damage detection method according to any embodiment of the present application.

[0025] In the technical solution of the embodiment of the present application, at least one decomposition vector is obtained by matrix decomposition of target signal data when a vehicle passes over the wheel or rail. Each decomposition vector is reconstructed based on the centroid frequency of each decomposition vector to obtain reconstructed signal data. The reconstruction process performs a large amount of denoising on the target signal data and retains the information of the damage signal to the maximum extent, thereby improving the efficiency of damage signal identification. The reconstructed signal data is subjected to a preset modal decomposition process to obtain at least one modal component. By performing modal decomposition on the reconstructed signal data, the reconstructed signal can be better distinguished into modes with different frequencies and energies, which also helps to improve the efficiency and accuracy of damage signal identification. The modal components are divided into two groups based on a preset frequency threshold, and the energy ratio of the two groups of modal components is calculated. Based on the energy ratio, the presence of a damage signal in the target signal data is determined, and the strength of the damage signal is identified by the energy ratio, thereby accurately identifying the damage signal in the target signal data and improving the accuracy of wheel or rail damage identification.

[0026] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0028] Figure 1 This is a flow chart of a wheel-rail damage detection method provided in accordance with the first embodiment of the present application;

[0029] Figure 2A is a schematic diagram of wheel-rail damage identification according to the second embodiment of the present application;

[0030] Figure 2B This is a time-frequency diagram of the damage signal and the vehicle noise shown in Example 2 of the present application;

[0031] Figure 2C This is a diagram showing the VMD decomposition results of the wheel-rail noise signal presented in Example 2 of the present application;

[0032] Figure 2D This is a diagram showing the VMD decomposition results of the acoustic emission signal under damage shown in Example 2 of the present application;

[0033] Figure 2EThis is a diagram showing the VMD decomposition result of the mixture of the passing vehicle noise and the damage signal shown in Example 2 of the present application;

[0034] Figure 2F is a schematic diagram of pure passing vehicle noise and mixed energy ratio according to Example 2 of the present application;

[0035] Figure 3 This is a structural diagram of a wheel-rail damage detection device provided according to the third embodiment of the present application;

[0036] Figure 4 It is a structural schematic diagram of an electronic device for implementing the wheel-rail damage detection method of an embodiment of the present application. DETAILED DESCRIPTION

[0037] In order to enable those skilled in the art to better understand the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of this application.

[0038] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in a sequence other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0039] Example 1

[0040] Figure 1 A flowchart of a wheel-rail damage detection method is provided for the first embodiment of the present application. This embodiment is applicable to the case of detecting damage to the track used by the train. The method can be executed by a wheel-rail damage detection device, which can be implemented in the form of hardware and / or software, and can be configured in an electronic device. Figure 1 As shown, the method includes:

[0041] S110: Acquire target signal data when a vehicle passes on the target wheel track.

[0042] The target wheel / rail may be the track to be inspected. It is understood that wheels and rails may become damaged due to physical factors such as compression and vibration after use. However, since the extent of damage cannot be directly observed with the naked eye, damage inspection is necessary to determine their health. The target signal data may be a vibration signal obtained by inspecting the target wheel / rail when a train passes over it. For example, an acoustic emission sensor may be installed on the target wheel / rail to continuously monitor and acquire vibration signals from the rail.

[0043] In an optional implementation manner, obtaining a target signal when a vehicle passes through the target wheel track in S110 may include:

[0044] S111. Acquire a signal to be processed obtained by performing acoustic emission detection on a target wheel-rail.

[0045] Acoustic emission sensors are used to collect all vibration signals generated on the target wheel-rail. It can be understood that regardless of whether a vehicle passes by the target wheel-rail, some signals to be processed can be collected. These may include white noise from the presence or absence of a vehicle, vibration signals from the presence of a vehicle, and damage signals reflected by damage to the target wheel-rail.

[0046] S112 , performing a preset Fourier transform on the signal to be processed to obtain a signal to be analyzed in the frequency domain.

[0047] Since the recording method for collecting these signals to be processed is to record them in a time sequence, for example, a certain number of data are recorded per second as a frame for subsequent processing, the time domain signal is converted into a frequency domain signal through Fourier transform for analysis. Of course, Fourier transform includes various forms, which can be selected by relevant technicians according to actual circumstances, such as fast Fourier transform, which is not limited in the embodiments of this application.

[0048] S113 , judging, based on the energy sum of the signals to be analyzed that meet the preset frequency range, whether the corresponding signals to be processed are collected when a vehicle passes on the target wheel-track.

[0049] The preset frequency range can be a threshold for determining whether a train has passed through the vibration frequency. For example, 50 kHz can be selected. The energy of the data below 50 kHz is summed to determine whether the energy corresponding to this data matches the energy generated by the wheel-rail vibration when a train passes over the target wheel-rail.

[0050] S114. All data corresponding to the signal to be processed that is collected when a vehicle passes on the target wheel track is used as target signal data.

[0051] Continuing with the previous example, all data corresponding to the signal to be processed whose energy and conformity with the vehicle passing condition are used as target signal data, that is, signal data used for subsequent damage detection.

[0052] Continuing with S110 and S120 , the target signal data is subjected to a matrix decomposition process in a preset manner to obtain at least one decomposition vector.

[0053] It should be noted that the signal data that can be detected when a train passes over the target wheel or rail includes multiple types, such as vibration signals generated by the passing train, damage signals reflected by wheel or rail damage, and white noise signals. These signals are mixed together and reflected in the target signal data. When analyzing whether the target wheel or rail is damaged, the target signal data must first be denoised. This noise removal process aims to eliminate other signals that affect wheel or rail damage detection (such as white noise signals and vibration signals generated by passing trains). Therefore, multiple sub-matrices are obtained through matrix decomposition, and each sub-matrix can be de-diagonalized to convert these sub-matrices into multiple decomposition vectors. These decomposition vectors can represent different data of the vibration signal. These decomposition vectors can classify different types of data (damage data, noise data), etc., thereby helping to eliminate the impact of noise on damage detection. Of course, the preset method used in the matrix decomposition process can be selected by relevant technicians based on a large number of experiments or actual situations. For example, the singular spectrum decomposition method can be used.

[0054] In an optional implementation, performing a preset matrix decomposition process on the target signal data in S120 to obtain at least one decomposition vector may include:

[0055] S121. Convert target signal data into time series data in chronological order.

[0056] The target signal data is acquired in a time series, so the data is converted into a time series according to the time series. For example, 8192 vibration signal data are acquired every second, and these 8192 data are combined into a frame of time series data.

[0057] S122: Construct the time series data into a trajectory matrix according to a preset method.

[0058] Constructing a trajectory matrix from time series data is a common technique in time series analysis. This technique is used to map one-dimensional time series data into a higher-dimensional space, thereby facilitating the discovery of underlying structures, patterns, or periodicities. Therefore, any method known in the relevant art can be used to construct the trajectory matrix, and this embodiment of the present application does not limit this.

[0059] S123. Perform singular spectrum decomposition on the trajectory matrix to obtain at least one decomposition vector.

[0060] Through singular spectrum decomposition, the trajectory matrix is ​​decomposed into multiple sub-matrices. Each sub-matrix can be decomposed into anti-diagonal averages, thereby converting the sub-matrix into a sequence of multiple column vectors, that is, decomposing the initial time series into multiple vectors.

[0061] In the above-mentioned implementation, the application of singular spectrum analysis to wheel-rail damage detection can provide favorable support for the subsequent removal of low-frequency wheel-rail noise. Further denoising processing can be performed on these decomposed vectors to improve the denoising efficiency, help retain the damage signal, and facilitate accurate identification of wheel-rail damage.

[0062] Continuing with S120 and S130 , each decomposition vector is reconstructed according to the centroid frequency of each decomposition vector to obtain reconstructed signal data.

[0063] The centroid frequency can be the weighted average frequency of the signal spectrum, reflecting the distribution of the main frequency band of the signal power spectrum, that is, the location of energy concentration in the signal spectrum. The centroid frequency of each decomposed vector is calculated separately, and vectors that meet the requirements are selected as the basis for reconstruction. These selected vectors are then recombined into reconstructed signal data. It can be understood that the reconstructed signal data has eliminated low-frequency interference in the original data, making it easier to detect damaged signals when testing the reconstructed signal data.

[0064] In an optional embodiment, reconstructing each decomposition vector according to the centroid frequency of each decomposition vector to obtain reconstructed signal data in S130 may include: adding each decomposition vector having a centroid frequency greater than a preset low-frequency threshold to obtain reconstructed signal data.

[0065] The preset low-frequency threshold can be a threshold for filtering out low-frequency noise. Decomposition vectors with a center-of-gravity frequency greater than the preset low-frequency threshold are retained, while decomposition vectors with a center-of-gravity frequency less than or equal to the preset low-frequency threshold are filtered out. Decomposition vectors with a center-of-gravity frequency greater than the preset low-frequency threshold are then summed to reconstruct reconstructed signal data. It can be understood that, compared to the target signal data, the reconstructed signal data filters out low-frequency noise and is more conducive to detecting whether the target wheel or rail is damaged.

[0066] Of course, the preset low-frequency threshold can be set by relevant technical personnel based on a large number of experiments or specific conditions of the wheel and rail. For example, it can be set to 100 kHz, and this embodiment of the present application does not limit this.

[0067] Continuing with S130 and S140 , a preset modal decomposition process is performed on the reconstructed signal data to obtain at least one modal component.

[0068] The preset modal decomposition process can be a variational modal decomposition method. In signal processing, variational modal decomposition is a signal decomposition estimation method. This method determines the frequency center and bandwidth of each component by iteratively searching for the optimal solution of the variational model during the decomposition process. This allows for adaptive frequency domain decomposition of the signal and effective separation of its components.

[0069] In an optional implementation, performing a preset modal decomposition process on the reconstructed signal data to obtain at least one modal component in S140 may include:

[0070] S141. Using the ratio of the kurtosis and envelope entropy of the reconstructed signal data as a fitness function.

[0071] Kurtosis is a numerical statistic that reflects the distribution characteristics of a variable, while envelope entropy reflects the uncertainty of the signal amplitude distribution. The ratio of kurtosis to envelope entropy is used as the fitness function. A larger ratio indicates a better decomposition effect and an easier identification of damaged signals.

[0072] S142: Load the fitness function into a preset selection particle swarm algorithm, and determine the optimal number of decomposition levels and the optimal penalty factor within a preset range of the number of decomposition levels and the range of the penalty factor.

[0073] Among them, the number of decomposition layers can be the target number of layers of variational mode decomposition, that is, how many modes the reconstructed signal data is divided into. The penalty factor is a key parameter in variational mode decomposition, which directly controls the constraint strength of the modal bandwidth during the decomposition process, and balances the compactness of the mode and the accuracy of signal reconstruction. Accordingly, the range of the number of decomposition layers and the range of the penalty factor can be pre-set by relevant technical personnel based on a large number of experiments or manual experience. For example, it can be set to a range of 4-8 decomposition layers and a range of 500-5000 penalty factors. The optimization algorithm of the selected particle swarm can determine the optimal parameters within the set range, that is, obtain the optimal number of decomposition layers and the optimal penalty factor. Of course, the difference between the selected particle swarm algorithm used in the embodiment of the present application and the related art is that the ratio of kurtosis and envelope entropy is used as the fitness function in this case.

[0074] S143. Perform variational modal decomposition on the reconstructed signal data according to the optimal number of decomposition layers and the optimal penalty factor to obtain at least one modal component.

[0075] Under the condition that the optimal number of decomposition layers and the optimal penalty factor are obtained in the aforementioned steps, a variational modal decomposition operation is performed on the reconstructed signal data to obtain multiple modal components for subsequent damage identification.

[0076] In the above-described embodiment, variational modal decomposition (VMD) yields multiple modal components. Compared to the unsteadiness and chaos of vibration signals, the decomposed modal components are clearer, resolving the problem of different frequency components being mixed within the same mode. Furthermore, using the ratio of kurtosis to envelope entropy as the fitness function in the particle swarm algorithm (PSO) optimizes the extraction of high-frequency transient components and improves the signal-to-noise ratio, significantly facilitating the subsequent identification of wheel-rail damage signals and enhancing the efficiency and accuracy of damage identification.

[0077] Continuing with S140 , S150 , each modal component is divided into two groups according to a preset frequency threshold, and an energy ratio of the two groups of modal components is calculated.

[0078] Among them, the preset frequency threshold can be the basis for grouping different modal components. It can be understood that each modal component has its own frequency domain characteristics. According to different frequency distributions, all modal components obtained by variational modal decomposition are grouped, and the energies of the two groups of modal components are statistically analyzed and ratio operations are performed to obtain the energy ratio of the two groups.

[0079] For example, the centroid frequency or center frequency of the modal components is used for determination. Modal components whose centroid frequency or center frequency exceeds a preset frequency threshold are grouped together, and modal components whose centroid frequency or center frequency does not exceed the preset frequency threshold are grouped together. The summed energies of the modal components in these two groups are calculated, and the ratio of the two summed energies is further calculated.

[0080] Of course, the preset frequency threshold can be further determined by technicians in related fields based on a large number of experiments or actual conditions. For example, it can be set to 150kHz, and the embodiments of this application do not make specific limitations.

[0081] In an optional implementation, the step of dividing the modal components into two groups according to a preset frequency threshold and calculating the energy ratio of the two groups of modal components in S150 may include:

[0082] S151. Calculate the center of gravity frequency of each modal component respectively.

[0083] Among them, the center of gravity frequency has been explained in the aforementioned embodiment and will not be repeated here. The center of gravity frequency can be calculated using any calculation method in the relevant technology, such as the frequency-weighted average calculation of each modal component with the power spectrum amplitude as the weight to obtain the center of gravity frequency of each modal component.

[0084] S152: Divide modal components whose center of gravity frequencies are greater than or equal to a preset frequency threshold into a first modal group, and divide modal components whose center of gravity frequencies are less than the preset frequency threshold into a second modal group.

[0085] Exemplarily, modal components with a center-of-gravity frequency greater than or equal to 150 kHz are divided into a first modal group, and modal components with a center-of-gravity frequency less than 150 kHz are divided into a second modal group.

[0086] S153 , respectively accumulating the energy values ​​of the modal components in the first modal group and the second modal group to obtain the first group total energy and the second group total energy.

[0087] The energy values ​​of all modal components in the first modal group are accumulated to obtain the first group total energy; the energy values ​​of all modal components in the second modal group are accumulated to obtain the second group total energy.

[0088] S154. The ratio of the total energy of the first group to the total energy of the second group is used as the energy ratio.

[0089] The ratio obtained by dividing the total energy of the first group by the total energy of the second group is taken as the energy ratio.

[0090] Continuing with S150 and S160 , it is determined whether a damaged signal exists in the target signal data based on the energy ratio.

[0091] Understandably, in actual applications, damage signals are primarily concentrated around 180kHz, while wheel-rail noise has energy up to 200kHz. Therefore, the total energy of the modal components above 150kHz (i.e., the first group of total energy) can more directly reflect the energy of the damage signal. The larger this energy ratio, the greater the proportion of the damage signal to the total signal. Of course, specific thresholds can also be set to determine the degree of damage to the target wheel-rail.

[0092] In an optional implementation, determining the presence of an impairment signal in the target signal data based on the energy ratio in S160 may include: determining the presence of an impairment signal in the target signal data in response to the energy ratio being greater than a preset impairment threshold.

[0093] The preset damage threshold may be a basis for judging the energy ratio. When the energy ratio is greater than the preset damage threshold, it may be considered that a damage signal exists in the target signal data, i.e., the target wheel-rail has damage characteristics.

[0094] For example, the preset damage threshold may be set to 2. When the energy ratio is greater than 2, it indicates that the damage signal occupies a large proportion, that is, the damage characteristics of the target wheel-rail are obvious.

[0095] In the technical solution of the above-mentioned embodiment of the present application, at least one decomposition vector is obtained by matrix decomposition of target signal data when a vehicle passes over the wheel or rail. Each decomposition vector is reconstructed based on the centroid frequency of each decomposition vector to obtain reconstructed signal data. The reconstruction process performs a large amount of denoising on the target signal data and maximizes the retention of damage signal information, thereby improving the efficiency of damage signal identification. The reconstructed signal data is subjected to a preset modal decomposition process to obtain at least one modal component. The modal decomposition of the reconstructed signal data can better distinguish the reconstructed signal into modes with different frequencies and energies, which also helps to improve the efficiency and accuracy of damage signal identification. The modal components are divided into two groups based on a preset frequency threshold, and the energy ratio of the two groups of modal components is calculated. The presence of a damage signal in the target signal data is determined based on the energy ratio, and the strength of the damage signal is identified based on the energy ratio, thereby accurately identifying the damage signal in the target signal data and improving the accuracy of wheel or rail damage identification.

[0096] Example 2

[0097] Figure 2A This is a schematic diagram of wheel-rail damage detection provided in Example 2 of this application. This example of the application is in the aforementioned examples and various implementation methods. Figure 2A As shown, the method includes:

[0098] In step (1), the hardware equipment used to detect wheel-rail damage is powered on and initialized. The hardware equipment mainly includes a collection part installed on the track (for example, an acoustic emission sensor) and a data processing part installed beside the track. The initial state is no vehicle passing, which is represented by the identifier 00. The first 0 represents the vehicle passing state in the previous second, and the second 0 represents the vehicle passing state in the current second. 00 does not start data storage. Therefore, it is conceivable that when the process changes from no vehicle on the track to vehicle passing on the track and then to no vehicle on the track, the identifier has a change process from 00 to 01, from 01 to 10, and finally back to 00.

[0099] In step (2), the vibration signal on the track is continuously monitored by the acquisition hardware equipment on the track, and 8192 points are taken per second for Fourier transform. After summing the energy of the signal below 50kHz, a vehicle passing judgment is performed. When it is judged that a vehicle is passing, the identifier becomes 1, and the vehicle passing status is represented as 01. The signal storage function is started, and the collected acoustic emission signal is transmitted to the trackside equipment for subsequent damage analysis; until the vehicle passing is completed and the vehicle passing status becomes 00, the data storage is terminated.

[0100] Step (3) The stored vehicle passing data is divided into one frame for every N points (N can be 8192), and the time series of this frame x=(x1,x2,…,x N) to perform singular spectrum and variable mode processing, and then identify damage through features. The specific steps of preprocessing using singular spectrum analysis (SSA) are as follows:

[0101] ①Construct the time series x into a trajectory matrix X: Set the window length to L, and the trajectory matrix is ​​as follows:

[0102]

[0103] Where K = N - L + 1, L is the set window length, and X is the trajectory matrix with a size of L × K dimensions.

[0104] ② Perform singular value decomposition (SVD) on the trajectory matrix X, which can be expressed as:

[0105] X=UΣV T

[0106] Where U∈RL×L is a left singular matrix, V∈RK×K is a right singular matrix, T is the transpose sign, Σ∈RL×K is a singular matrix, and the non-zero elements in Σ are singular values.

[0107] ③ Group the matrix X, generally according to the number of non-zero singular values. If the number of non-zero singular values ​​is m, then X can be divided into m linearly independent sub-matrices, that is:

[0108] X=X1+X2+X3...+X m

[0109] The i-th submatrix X i It can be expressed as:

[0110]

[0111] where u i 、v i are the i-th column vectors of matrices U and V, σ i are the singular values ​​corresponding to Σ.

[0112] ④For each submatrix X i Perform anti-diagonal averaging, that is, convert the matrix into a sequence y with a length of N = L + K-1 i ,i=1,2,…,m.

[0113] In step (4), after decomposing the original sequence into m vectors, in order to filter out low-frequency interference within 100kHz, each sequence y is calculated separately. i The center frequency fc of the signal with fc greater than 100kHz is reconstructed and the reconstructed signal is recorded as x re .

[0114] Step (5), parameter optimization: Perform variational mode decomposition (VMD) on the reconstructed signal. Two parameter factors need to be set in VMD decomposition: the number of decomposition layers k and the penalty factor α. The two parameters are determined as follows:

[0115] ① According to the complexity of the signal and the empirical value, set the range of k to 4-8; set α to 500-5000;

[0116] ② Due to the high-frequency characteristics of the damage signal and its impact signal, the ratio of the two indicators of kurtosis and envelope entropy is used as the fitness function. The larger the ratio, the better the decomposition effect and the easier it is to identify the damage signal.

[0117] ③ Select the Particle Swarm Optimization (PSO) algorithm to globally find the optimal values ​​for the number of decomposition levels k and the penalty factor α. To balance computational efficiency and search capability, the PSO algorithm can be set to a swarm size of 30 and a maximum number of iterations of 100.

[0118] ④According to the above parameters, the optimization process is executed and the optimal parameter is finally determined as k opt , α opt .

[0119] Step (6), for the signal x re Perform VMD decomposition and get k opt The new modal components are selected. According to the actual data, the damage signal is concentrated at 180kHz, while the wheel-rail noise has energy within 200kHz. The signals are divided into two groups according to the center frequency (or center of gravity frequency) of the new modal components. The frequency bands of the first k1 modes are above 150kHz, and the frequency bands of the last k modes are above 150kHz. opt The frequency band of the k1 modes is below 150 kHz. The energy ratio of the two groups of signals is calculated as R = E1 / E2. The threshold of this energy ratio is set to Rsh. If R is greater than the damage threshold Rsh, it is determined that a damaged signal exists in this segment of data.

[0120] In step (7), steps (3) to (6) are repeated for the data stored during the vehicle passing, and the number of damage signals detected during this vehicle passing is counted.

[0121] In addition to the above method steps, relevant experimental verification was carried out to verify the feasibility and accuracy of the method described in the embodiment of this application. In order to verify the effectiveness of the algorithm under strong noise interference, actual train travel data on a heavy-load line was used as the real environmental noise, and a pressure test was performed on a sample made of the same material as the rail to obtain the actual damage signal. The simulated damage has similar characteristics to the actual rail crack, which ensures that the proposed detection method is effective. The damage signal and the time-frequency diagram of the passing noise are shown in Figure 2. Figure 2B As shown, the passing vehicle signal and the actual damage signal can be distinguished intuitively, and the sampling rate is fs = 1MHz.

[0122] The above signal is framed. Each frame of data has N=8192, with a total of 50 sample points. The 50 samples of vehicle noise and vehicle noise mixed with damage signals are decomposed into 10 components by singular spectrum decomposition. L=10 is taken. The components are reconstructed into x according to their frequency components. re Then x re Perform VMD decomposition, Figure 2C 、 Figure 2D 、 Figure 2E These are the spectral components of three typical modes after VMD decomposition. Figure 3 The decomposition result under pure passing vehicle noise, based on which R=1.5 is calculated. Figure 2D Figure 2 is the decomposition result of the damage signal. It can be seen that the 180kHz damage signal is well extracted, and R=4.28 is calculated. Figure 2E The decomposition result after mixing the two signals is shown in Figure 2. The energy ratio R = 2.68. Based on experience, Rsh = 2 is set as the threshold for determining whether the signal is damaged. Figure 2F The samples in the middle circle represent pure traffic noise, and the samples marked with asterisks represent mixed traffic noise. This shows that by combining singular spectrum analysis with VMD decomposition, damage signals can be extracted from the decomposed modal features.

[0123] Example 3

[0124] Figure 3 This is a structural schematic diagram of a wheel-rail damage detection device provided in Example 3 of the present application.

[0125] like Figure 3 As shown, the device 300 includes:

[0126] The signal acquisition module 310 is used to acquire target signal data when a vehicle passes on the target wheel track;

[0127] The matrix decomposition module 320 is used to perform a matrix decomposition process on the target signal data in a preset manner to obtain at least one decomposition vector;

[0128] A signal reconstruction module 330 is used to reconstruct each decomposition vector according to the centroid frequency of each decomposition vector to obtain reconstructed signal data;

[0129] A modal decomposition module 340 is configured to perform a preset modal decomposition process on the reconstructed signal data to obtain at least one modal component;

[0130] an energy comparison module 350 for dividing each modal component into two groups according to a preset frequency threshold and calculating an energy ratio of the two groups of modal components;

[0131] The impairment determination module 360 ​​is configured to determine whether an impairment signal exists in the target signal data based on the energy ratio.

[0132] In the technical solution of the above-mentioned embodiment of the present application, at least one decomposition vector is obtained by matrix decomposition of target signal data when a vehicle passes over the wheel or rail. Each decomposition vector is reconstructed based on the centroid frequency of each decomposition vector to obtain reconstructed signal data. The reconstruction process performs a large amount of denoising on the target signal data and maximizes the retention of damage signal information, thereby improving the efficiency of damage signal identification. The reconstructed signal data is subjected to a preset modal decomposition process to obtain at least one modal component. The modal decomposition of the reconstructed signal data can better distinguish the reconstructed signal into modes with different frequencies and energies, which also helps to improve the efficiency and accuracy of damage signal identification. The modal components are divided into two groups based on a preset frequency threshold, and the energy ratio of the two groups of modal components is calculated. The presence of a damage signal in the target signal data is determined based on the energy ratio, and the strength of the damage signal is identified based on the energy ratio, thereby accurately identifying the damage signal in the target signal data and improving the accuracy of wheel or rail damage identification.

[0133] In an optional implementation, the modal decomposition module 340 may include:

[0134] a fitness function determination unit, configured to use the ratio of the kurtosis and envelope entropy of the reconstructed signal data as a fitness function;

[0135] An optimal parameter determination unit is used to load the fitness function into a preset selection particle swarm algorithm, and determine the optimal number of decomposition layers and the optimal penalty factor within a preset range of the number of decomposition layers and the range of the penalty factor;

[0136] The modal component determination unit is used to perform variational modal decomposition on the reconstructed signal data according to the optimal decomposition layer number and the optimal penalty factor to obtain at least one modal component.

[0137] In an optional implementation, the matrix decomposition module 320 may include:

[0138] A time series conversion unit, used to convert target signal data into time series data in time sequence;

[0139] A matrix construction unit, used to construct the time series data into a trajectory matrix according to a preset method;

[0140] The singular spectrum decomposition unit is used to perform singular spectrum decomposition on the trajectory matrix to obtain at least one decomposition vector.

[0141] In an optional implementation, the signal reconstruction module 330 may be specifically configured to:

[0142] The decomposition vectors whose centroid frequencies are greater than a preset low-frequency threshold are added together to obtain reconstructed signal data.

[0143] In an optional implementation, the energy comparison module 350 may include:

[0144] A center of gravity frequency calculation unit, used to calculate the center of gravity frequency of each modal component respectively;

[0145] a modal grouping unit, configured to group modal components having a center of gravity frequency greater than or equal to a preset frequency threshold into a first modal group, and group modal components having a center of gravity frequency less than the preset frequency threshold into a second modal group;

[0146] an energy accumulation unit, configured to accumulate the energy values ​​of the modal components in the first modal group and the second modal group respectively to obtain the total energy of the first group and the total energy of the second group;

[0147] The energy ratio calculation unit is used to take the ratio of the total energy of the first group to the total energy of the second group as the energy ratio.

[0148] In an optional implementation, the damage determination module 360 ​​may be specifically configured to:

[0149] In response to the energy ratio being greater than a preset impairment threshold, it is determined that an impairment signal exists in the target signal data.

[0150] In an optional implementation, the signal acquisition module 310 may include:

[0151] a signal-to-be-processed acquisition unit, configured to acquire a signal-to-be-processed obtained by performing acoustic emission detection on a target wheel-rail;

[0152] A Fourier transform unit, configured to perform a preset Fourier transform on the signal to be processed to obtain a signal to be analyzed in the frequency domain;

[0153] An energy judgment unit is used to judge whether the corresponding signal to be processed is collected when a vehicle passes on the target wheel track based on the energy sum of the signal to be analyzed that meets the preset frequency range;

[0154] The target signal determination unit is used to use all data corresponding to the to-be-processed signal collected when a vehicle passes on the target wheel track as target signal data.

[0155] The wheel-rail damage detection device provided in the embodiments of the present application can execute the wheel-rail damage detection method provided in any embodiment of the present application, and has the corresponding functional modules and beneficial effects for executing each wheel-rail damage detection method.

[0156] Example 4

[0157] Figure 4 A schematic diagram of the structure of an electronic device 10 that can be used to implement an embodiment of the present application is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present application described and / or required herein.

[0158] like Figure 4 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., which is communicatively connected to the at least one processor 11. The memory stores a computer program that can be executed by the at least one processor. The processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 into the random access memory (RAM) 13. Various programs and data required for the operation of the electronic device 10 can also be stored in the RAM 13. The processor 11, ROM 12, and RAM 13 are connected to each other via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0159] Multiple components in the electronic device 10 are connected to the I / O interface 15, including an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a magnetic disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.

[0160] The processor 11 can be any general-purpose and / or specialized processing component with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various specialized artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any other suitable processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as the wheel-rail damage detection method.

[0161] In some embodiments, the wheel / rail damage detection method may be implemented as a computer program tangibly embodied in a computer-readable storage medium, such as the storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the wheel / rail damage detection method described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to execute the wheel / rail damage detection method in any other appropriate manner (e.g., via firmware).

[0162] Various embodiments of the systems and techniques described herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), system-on-chip systems (SOCs), programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include being implemented in one or more computer programs that are executable and / or interpreted on a programmable system that includes at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0163] Computer programs for implementing the methods of the present application may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when the computer program is executed by the processor, the functions / operations specified in the flowcharts and / or block diagrams are implemented. The computer program may be executed entirely on the machine, partially on the machine, as a stand-alone software package, partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0164] In the context of the present application, a computer-readable storage medium can be a tangible medium that can contain or store a computer program for use by an instruction execution system, device or equipment or used in combination with an instruction execution system, device or equipment. A computer-readable storage medium can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium can be a machine-readable signal medium. A more specific example of a machine-readable storage medium can include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0165] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0166] The systems and techniques described herein can be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0167] A computing system may include clients and servers. The clients and servers are typically remote from each other and typically interact via a communication network. This client-server relationship arises through computer programs running on the respective computers, creating a client-server relationship. The server may be a cloud server, also known as a cloud computing server or cloud host. This server is a hosting product within the cloud computing service ecosystem that addresses the management difficulties and limited scalability of traditional physical hosting and VPS services.

[0168] The present application also discloses a computer program product comprising a computer program that, when executed by a processor, implements the wheel-rail damage detection method provided in any of the embodiments of the present application. This program product shares the same inventive concept as the wheel-rail damage detection method disclosed in each embodiment of the present application and is therefore not further described here.

[0169] It should be understood that the various forms of the processes shown above can be used to reorder, add, or delete steps. For example, the steps described in this application can be performed in parallel, sequentially, or in a different order, as long as the desired results of the technical solution of this application can be achieved. This is not limited herein.

[0170] The above specific embodiments do not constitute a limitation on the scope of protection of this application. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions may be made based on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application shall be included within the scope of protection of this application.

Claims

1. A wheel-rail damage detection method, characterized in that: include: Obtain target signal data when a vehicle passes on the target wheel track; Performing a matrix decomposition process on the target signal data in a preset manner to obtain at least one decomposition vector; Reconstructing each of the decomposition vectors according to the centroid frequency of each of the decomposition vectors to obtain reconstructed signal data; Performing a preset modal decomposition process on the reconstructed signal data to obtain at least one modal component; Dividing each of the modal components into two groups according to a preset frequency threshold, and calculating an energy ratio of the two groups of modal components; According to the energy ratio, it is determined that a damaged signal exists in the target signal data.

2. The method according to claim 1, characterized in that The performing a preset modal decomposition process on the reconstructed signal data to obtain at least one modal component includes: Using the ratio of the kurtosis and envelope entropy of the reconstructed signal data as a fitness function; The fitness function is loaded into a preset selection particle swarm algorithm, and an optimal number of decomposition levels and an optimal penalty factor are determined within a preset range of decomposition levels and a preset range of penalty factors; According to the optimal decomposition layer number and the optimal penalty factor, variational modal decomposition is performed on the reconstructed signal data to obtain at least one modal component.

3. The method according to claim 1, characterized in that The step of performing matrix decomposition processing on the target signal data in a preset manner to obtain at least one decomposition vector includes: Converting the target signal data into time series data in chronological order; Constructing the time series data into a trajectory matrix according to a preset method; Performing singular spectrum decomposition on the trajectory matrix to obtain at least one decomposition vector.

4. The method according to claim 1, wherein The reconstructing each of the decomposition vectors according to the centroid frequency of each of the decomposition vectors to obtain reconstructed signal data includes: The decomposition vectors having the center of gravity frequency greater than a preset low-frequency threshold are added together to obtain the reconstructed signal data.

5. The method according to claim 1, characterized in that The dividing the modal components into two groups according to a preset frequency threshold and calculating the energy ratio of the two groups of modal components includes: Calculating the center of gravity frequency of each modal component respectively; Dividing the modal components whose center of gravity frequency is greater than or equal to the preset frequency threshold into a first modal group, and dividing the modal components whose center of gravity frequency is less than the preset frequency threshold into a second modal group; Accumulating the energy values ​​of the modal components in the first modal group and the second modal group respectively to obtain a first group total energy and a second group total energy; The ratio of the total energy of the first group to the total energy of the second group is used as the energy ratio.

6. The method according to claim 1, characterized in that Determining, based on the energy ratio, whether a damaged signal exists in the target signal data includes: In response to the energy ratio being greater than a preset impairment threshold, it is determined that an impairment signal exists in the target signal data.

7. The method according to claim 1, characterized in that The acquiring of the target signal when a vehicle passes on the target wheel track comprises: Acquiring a signal to be processed obtained by performing acoustic emission detection on the target wheel-rail; Performing a preset Fourier transform on the signal to be processed to obtain a signal to be analyzed in the frequency domain; According to the energy sum of the signal to be analyzed that meets the preset frequency range, it is determined whether the corresponding signal to be processed is collected when a vehicle passes through the target wheel track; All data corresponding to the signal to be processed that is collected when a vehicle passes on the target wheel track is used as the target signal data.

8. A wheel-rail damage detection device, characterized in that: include: A signal acquisition module is used to obtain target signal data when a vehicle passes on the target wheel track; a matrix decomposition module, configured to perform a matrix decomposition process on the target signal data in a preset manner to obtain at least one decomposition vector; A signal reconstruction module, configured to reconstruct each of the decomposition vectors according to the centroid frequency of each of the decomposition vectors to obtain reconstructed signal data; a modal decomposition module, configured to perform a preset modal decomposition process on the reconstructed signal data to obtain at least one modal component; an energy comparison module, configured to divide each of the modal components into two groups according to a preset frequency threshold, and calculate an energy ratio of the two groups of modal components; The impairment determination module is configured to determine whether an impairment signal exists in the target signal data based on the energy ratio.

9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the wheel-rail damage detection method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the wheel-rail damage detection method according to any one of claims 1 to 7 when executed.

Images